Homozygous loss of putative "tumorigenicity suppressor" genes seems to be etiologically important in malignant transformation. Molecular marker studies of different cancers reveal a high incidence of homozygous deletions that may include sites of suppressor genes. Cytogenetic studies also reveal nonrandom losses of whole or segments of chromosomes as well as gains and often tumor type-specific chromosome rearrangements. That specific chromosomes or often only certain chromosome regions are involved in many different neoplasias suggests that there may be genes or gene clusters at these sites whose loss or alteration can lead to transformation in different tissues. On the other hand, different alterations are found in some tumors that are of the same histopathologic type. It is not clear how extensive such allelic losses and other nonrandom alterations are in some common cancers, such as cervical carcinomas, or to what degree these are directly related to cancer development. In some cancers, the number of allelic losses or the presence of specific chromosome rearrangements or duplications seem to relate to prognostic factors or sometimes correlate with the response of a tumor to therapy. The aims of this study of cervical carcinomas, one of the most frequent cancers in women, are: 1) To obtain a map (or allelotype) as detailed as practical of chromosome regions most commonly lost, duplicated, or altered in these cancers. 2) To obtain karyotypic information on concurrent chromosome changes. 3) To determine if any of these changes are prognostic indicators and if they might assist in the choice of therapeutic procedures. To establish each allelotype, DNA from carcinoma cells will be compared with DNA from the patient's normal constitutive cells on Southern blots. Probes will be used for regions that are adequately polymorphic to allow distinction of the two corresponding parental alleles in a high portion of DNA samples of normal cells. These probes detect reliably mapped regions of restriction fragment length polymorphisms (RFLPs) or of variable number of tandem repeats (VNTRs). A panel of 39 such probes will be used, each one specific to one of the major auto-some arms, to estimate the incidence of allelic losses, duplications, or alterations in the carcinoma cells. The comparison of the allelotype with the banded karyotype of the same tumor, which has not been done in most such studies, should allow not only estimation of the incidence of allelic alterations, but also indicate the extent of concurrent nonrandom chromosome changes. It is expected that this information when correlated with results of other studies will provide some insight into the molecular pathogenesis of cervical carcinomas. Finally, practical benefits could result if specific allelic or chromosome changes have prognostic value or can guide therapeutic choices.